Microsoft Chairman and CEO Satya Nadella warned on Tuesday that the artificial intelligence (AI) boom risks turning into a bubble if its benefits fail to spread beyond technology companies and into the real economy, arguing that AI’s long-term success depends on broad diffusion across industries, countries and workers.
“If all we are talking about are the tech firms, if all we talk about is what’s happening on the technology side, then that’s a bubble,” Nadella said in a conversation with BlackRock Chairman and CEO Larry Fink at the World Economic Forum (WEF) 2026 in Davos. “For this not to be a bubble, by definition it requires that the benefits of this are much more evenly spread.”
Nadella said AI should be judged not by the sophistication of models but by whether it delivers tangible outcomes such as faster drug development, improved healthcare, better education and higher productivity across sectors. “If these tokens are not improving health outcomes, education outcomes, public sector efficiency and private sector competitiveness, then this makes no sense,” he said.
The Microsoft chief framed AI as the latest step in a decades-long arc of computing, comparable to the advent of the personal computer, the internet and cloud computing. “Irrespective of the paradigm, it has been one continuous arc of saying let’s make better sense of this world by reasoning about it in digital form,” he said, adding that AI could ultimately have an even bigger impact.
AI, Power And The Human Agency
Nadella introduced the idea of AI output as a new economic commodity, repeatedly referring to “tokens,” units of AI computation, and likening them to electricity. “Tokens are going to be diffused all around the world, just like electricity,” he said, arguing that future economic growth would be closely linked to how cheaply and efficiently countries can produce and use them.
He said the economics of AI were already moving rapidly in that direction, noting that token prices were falling sharply as computing efficiency improved. “You have a commodity whose prices are just going to monotonically come down in a pretty fast curve,” Nadella said.
However, he cautioned that diffusion would not happen automatically. Access to energy, modern power grids, data centre infrastructure and skills would determine which countries and companies benefit most. “We will quickly lose the social permission to take something like energy, which is a scarce resource, and use it to generate these tokens if the outcomes are not there,” he said.
On jobs and productivity, Nadella pushed back against fears that AI would simply replace workers, describing it instead as a “cognitive amplifier” that changes how work is done. He pointed to software development as an early example, where AI has evolved from code completion tools to autonomous agents that can handle entire projects, while still leaving humans in control.
“It’s not about somehow living outside of the realm of human agency,” he said. “The levels of abstraction are going to change, but people still have agency.”
Nadella said organisations would need to redesign workflows and structures to realise productivity gains, warning that simply adding AI tools without changing how work is organised would limit impact. He described how AI tools such as Microsoft’s Copilot were flattening information flows inside companies, replacing traditional hierarchies of briefings and approvals. “It flattens the entire information flow,” he said.
He also talked about a “barbell effect” in adoption, with startups able to build around AI from scratch while large incumbents struggled with change management, even as they retained advantages in data, scale and customer relationships. “Neither side can just coast,” Nadella said.
On the global picture, Nadella said AI had the potential to spread more evenly than earlier technologies, including in emerging markets, because models and their outputs were accessible almost everywhere. But he stressed that capital investment and supportive policy environments remained essential. “As long as there’s an environment which attracts the capital investment and the demand is there, diffusion can happen,” he said.
On Europe And Sovereignty
Turning to Europe, Nadella warned against focusing too narrowly on technological sovereignty at the expense of competitiveness. European firms, he said, must ensure their products and services remain globally competitive. “You’re only going to be competitive if the products coming out of Europe are globally competitive,” he said.
Nadella also reframed the debate on AI sovereignty, arguing that the most critical issue in the coming year would be the sovereignty of firms rather than just nations. “The topic that’s least talked about, but I feel will be most talked about this calendar year, will be the sovereignty of a firm,” he said, warning that companies risk losing control over their competitive advantage if they fail to embed their own knowledge into AI systems they control.
“If your firm is not able to embed the tacit knowledge of the firm in a set of weights in a model that you control, by definition you have no sovereignty,” Nadella said, adding that focusing narrowly on data localisation misses the point.
“The data centre where it runs is the least important thing,” he said, cautioning that otherwise companies face “a one-way enterprise value transfer to some model company somewhere” if they rely entirely on external AI models rather than building and orchestrating their own. |
